Abstract: Far-field long-period harmonic-like ground motion characterized with long-period acceleration harmonic-like pulses in the late stage of vibration has adverse effects on the long-period structures. However, there still lacks of far-field ground motion records that could be directly used for seismic analysis due to limited records of ground motions. A synthetic method for far-filed harmonic-like ground motion based on Ensemble Empirical Mode Decomposition Method (EEMD) was proposed to solve the problem of shortage of available far-field ground motions, and its accuracy and feasibility was verified. The recorded far-field long-period harmonic-like ground motions were decomposed through EEMD to obtain the main-pulse composition and attenuation of low-frequency component, which were fitted and reconstructed to generate harmonic-like main-pulse component. Then the identification and simplification of characteristic parameters for the synthetic harmonic-like main-pulse components were conducted to obtain the main-pulse velocity model, which composed of harmonic-like main-pulse function with attenuation function. The synthetic method for far-filed harmonic-like ground motion was proposed based on combination of the main-pulse velocity model and high-frequency IMF components. In addition, the synthesis of three arbitrary harmonic-like ground motions according to selected characteristics parameters was carried out, and the comparisons with the original far-field long-period harmonic-like ground motions was conducted. The results demonstrated that the low-frequency component of original ground motion could be effectively extracted with EEMD and the velocity-time history obtained from the proposed main-pulse velocity model fitted well with the velocity-time history of the original ground motion. The proposed synthetic method could generate artificial harmonic-like ground motion that retained the non-stationary characteristics of the far-field harmonic-like ground motion, and fit well with the original ground motion.